Research On Computing Offloading And Resource Management Algorithm For Space-air-ground Internet Of Remote Things Networks

In the Internet of remote things(IoRT),smart devices are usually densely deployed in vast,remote areas without access to ground network services.Therefore,the IoRT network needs to break through the limitations of space and time to meet the needs of various technical applications(such as earth observation,military warfare,ecological environment monitoring,etc.).Compared with ground networks,satellite networks and air networks can provide greater coverage and have been used to provide Internet services to islands,remote mountainous areas and disaster areas.Therefore,a new network architecture--space-air-ground integrated networks(SAGIN)has attracted extensive attention.Among them,the satellite network,as a supplement and expansion of the ground network,provides an important solution to achieve the comprehensive coverage of the Internet of things(IoT)services.In the case of limited or no infrastructure such as base stations,unmanned aerial vehicles(UAV)can assist edge computing to provide computing offloading opportunities for smart devices,or it can be equipped with wireless energy transmission modules for smart devices to perform wireless energy transmission,and it can be deployed dynamically to ensure the normal operation of smart devices.This thesis is oriented to the computing tasks of vertical applications in the IoRT network,and considers the remote and vast geographic environment of the IoRT network,the size of smart devices is limited,the computing power is low,and the battery life is limited and difficult to replace.UAV is used to provide flexible and reliable computing services,satellites provide seamless coverage and always-on cloud computing services,thus forming a SAG-IoRT architecture.In view of the rich communication resources and diversified business types in the SAG-IoRT network architecture,this thesis focuses on the computing offloading and resource management algorithms of the SAG-IoRT network.The main work of this thesis is as follows:Firstly,considering that IoRT is located in remote areas where ground infrastructure is scarce and smart devices have limited battery life and computing capacity,the problem of minimizing the energy consumption of computing offloading in SAG-IoRT network is studied.This problem is modeled as a mixed integer nonlinear programming(MINLP)problem for the joint optimization of device scheduling,task assignment,transmission bit assignment and UAV trajectory planning in SAG-IoRT networks.Since the problem cannot be solved directly,this thesis divided it into three sub-problems for processing,solved the three sub-problems by variable relaxation and Lagrange dual decomposition methods,then iterated the three sub-problems continuously to obtain the minimum total system energy consumption.Finally,the convergence and effectiveness of the proposed algorithm are verified by simulation.The simulation results prove that the proposed algorithm can reduce system energy consumption to a certain extent.Secondly,in view of the limited battery life of IoRT and the difficulty of replacement,and considering that the SAG-IoRT network operates in a highly complex environment,the computing performance optimization of SAG-IoRT network based on UAV-assisted wireless power transmission(WPT)is studied.It is modeled as a stochastic resource management problem for the joint optimization of computing power,power control and UAV trajectory planning,aiming to ensure the stability and sustainability of the SAG-IoRT network and maximize the long-term time-average computational performance of the system.Since the proposed problem is a nonlinear stochastic optimization problem,this thesis first uses the Lyapunov optimization theory to decouple the stochastic problem into three sub-problems optimized under a single slot,and alternately iterate the three sub-problems to obtain the optimal under a single slot solution.On this thesis,we propose a stochastic resource management algorithm combining computing power,power control and UAV trajectory planning to effectively optimize system computing performance in an online manner.Finally,the sustainability and stability of the proposed algorithm are verified by simulation.Compared with the set baseline,the system performance of the proposed algorithm has superiority under certain circumstances.Finally,considering the complexity of SAG-IoRT network,the diversity of resources and the diversity of services,this thesis proposes a SAG-IoRT architecture based on network slicing,which is divided into horizontal slices composed of space layer,air layer and ground layer,and vertical slices of different service types.On this thesis,a joint optimization problem of device connection,transmission power,central processing unit(CPU)cycle frequency allocation and UAV deployment is proposed to maximize operator revenue.This resource management problem is a non-convex optimization problem.In order to make the problem easy to solve,this thesis divides it into three sub-problems and solved by variable relaxation,variable substitution and successive convex approximation(SCA).Then,the three sub-problems obtain the maximum revenue of the operator through alternate iterations.Finally,the resource management algorithm is simulated and verified.The numerical simulation of the proposed algorithm shows the effectiveness of the proposed algorithm.Compared with other schemes,the proposed algorithm can significantly improve the operator revenue of SAG-IoRT slice network.